CHAPTER 2

Microbial Growth Kinetics

VV1JA"~'~ in Chapter 1, fermentations may be car-

Lag Log or Stationary
batch, continuous and fed-batch processes. phase exponential phase
opi~ration is, to a large extent, dictated by phase
product being produced. This chapter will
kinetics and applications of batch, contin-
teCj-batc:n processes.
c
o
.;:;
BATCH CULTURE g
c
'"c
C,)

culture is a closed culture system which con- oC,)

initial, limited amount of nutrient. The inocu-
will pass through a number of phases, as
in Fig. 2.1. After inoculation there is a
which it appears that no growth takes
period is referred to as the lag phase and
considered as a time of adaptation. In a
CCllnlrlen;ial process the length of the lag phase should
as much as possible and this may be
llctlie,red by using a suitable inoculum, as described in
6. Following a period during which the growth
cells gradually increases, the cells grow at a
COllstamt, maximum, rate and this period is known as
or exponential, phase. The exponential phase
described by the equation:
dx/dt = /LX (2.1)
X is the concentration of microbial biomass,
t is time, in hours
/L is the specific growth rate, in hours -1.
mt(~gr:ati(lll equation (2.1) gives:
(2.2)
Xois the original biomass concentration,
is the biomass concentration after the time
interval, t hours,
e is the base of the natural logarithm.
'"
'"
'"E
o
:0
E
Time
FiG. 2.1. Growth of a typical microbial culture in batch conditions.
On taking natural logarithms, equation (2.2) becomes:
In XtIn Xo + /Lt.
= (2.3)
Thus, a plot of the natural logarithm of biomass con-
centration against time should yield a straight line, the
slope of which would equal JL. During the exponential
phase nutrients are in excess and the organism is
growing at its maximum specific growth rate, /Lrnax' for
the prevailing conditions. Typical values of /L rnax for a
range of micro-organisms are given in Table 2.1.
It is easy to visualize the exponential growth of
single celled organisms which replicate by binary fis-
sion. Indeed, animal and plant cells in suspension cul-
ture will behave very similarly to unicellular micro-
organisms (Griffiths, 1986; Petersen and Alfermann,
1993). However, it is more difficult to appreciate that

13
Principles of Fermentation Technology, 2nd Edn.

TABLE 2.1. Some representative values of /Lmax (obtained under the It is possible for new pellets to be generated by
conditions specified in the original reference) for a range of organisms
fragmentation of old pellets and, thus, the behaviour
a pelleted culture may be intermediate between expo
Organism Reference
nential and cube root growth.
Vibrio natriegens 4.24Eagon (1961)
Whether the organism is unicellular or mycelial th
Methylomonas methanolytica 0.53Dostalek et al. foregoing equations predict that growth will continu
(1972) indefinitely. However, growth results in the consum
Aspergillus nidulans 0.36 Trinci (1969) tion of nutrients and the excretion of
Penicillium chrysogenum 0.12 Trinci (1969)
products; events which influence the growth of
Fusarium graminearum Schwabe 0.28 Trinci (1992)
Plant cells in suspension 0.01-0.046 Petersen and organism. Thus, after a certain time the growth rate
culture Alfermann (1993) the culture decreases until growth ceases. The
Animal cells 0.01-0.05 Lavery (1990) tion of growth may be due to the depletion of
essential nutrient in the medium (substrate linlit'ltioln).
the accumulation of some autotoxic product of
mycelial organisms which show apical growth also grow organism in the medium (toxin limitation) or a
exponentially. Plomley (1959) was the first to suggest nation of the two.
that filamentous fungi have a 'growth unit' which is The nature of the limitation of growth may be
replicated at a constant rate and is composed of the plored by growing the organism in the presence
apex of the hypha and a short length of supporting range of substrate concentrations and plotting the
hypha. Trinci (1974) demonstrated that the total hyphal mass concentration at stationary phase against the
length of a mycelium and the number of tips increased tial substrate concentration, as shown in Fig. 2.2.
exponentially at approximately the same rate. Thus, Fig. 2.2 it may be seen that over the zone A to B
when the volume of the hyphal growth unit exceeds a increase in initial substrate concentration gives a
critical volume a new branch, and hence, a new growing portional increase in the biomass produced at
point, is initiated. This is equivalent to the division of a ary phase. The situation may be described by
single cell when the cell reaches a critical volume. equation:
Hence, the rate of increase in hyphal mass, total length
and number of tips is dictated by the specific growth X = Y(SR - s)
rate and:
dx/dt = /-LX, where is the concentration of biomass prcldulce,d,
X
Y is the yield factor (g biomass produced
dH/dt = /-LH, substrate consumed),
dA/dt = /-LA SR is the initial substrate concentration, and
s is the residual substrate concentration.
where H is total hyphal length and A is the number of
Over the zone A to B in Fig. 2.2, s equals zero at
growing tips.
point of cessation of growth. Thus, equation (2.4)
In submerged culture (shake flask or fermented a
be used to predict the biomass which may be prclduced
mycelial organism may grow as dispersed hyphal frag-
ments or as pellets (see also Chapters 6 and 9). The
growth of pellets will be exponential until the density Ie 0 1
of the pellet resvlts in diffusion limitation. Under such
c:
___rl _ _- - - - 1 1

limitation the central biomass of the pellet will not

receive a supply of nutrients, nor will potentially toxic ...co co'"
.2 OJ I
I
products diffuse out. Thus, the growth of the pellet ~-E. I
"'>
<J '-' 1
proceeds from the outer shell of biomass which is the c: co
o c: I
<J 0
actively growing zone and was described by Pirt (1975) ~ ..j:i I
co co
as: Et> 1
0'"
.- co I
M I / 3 = kt + MJ / 3 a:l I

where M o and M are the mycelium mass at time 0 and
t, respectively. Thus, a plot of the cube root of mycelial
mass against time will give a straight line, the slope of
which equals k.
Initial substrate concentration
FIG. 2.2. The effect of initial substrate eoncentration on the
mass eoneentration at the onset of stationary phase, in
culture.

14

of substrate. Over the zone C to
initial substrate concentration
Of(JPc)rtional increase in biomass. This
the exhaustion of another subs-
aCC;UlTml<lticm of toxic products. Over the
utiliz,lticm of the substrate is deleteri-
the accumulating toxins or the
qri"H,pr substrate.
is a measure of the efficiency of
one substrate into biomass and it
predict the substrate concentration
pflJdtlCe a certain biomass concentration.
iInpOlrtaJllt to appreciate that Y is not a
vary according to growth rate, pH,
limiting substrate and the concentra-
sul)strat1es in excess.
in growth rate and the cessation of
the depletion of substrate, may be
relationship between /.L and the resid-
IVtncurnmmg substrate, represented in equation
2.3 (Monad, 1942):
(2.5)
residual substrate concentration,
substrate utilization constant, numeri-
equal to substrate concentration when
/.Lmax and is a measure of the
of the organism for its substrate.
A to B in Fig. 2.3 is equivalent to the
phase in batch culture where substrate
;~Jitra.tion is in excess and growth is at /.L max ' The
in Fig. 2.3 is equivalent to the deceleration
culture where the growth of the organ-
re~iUllted in the depletion of substrate to a
IA BI
I middle exponential phase.
limiting substrate concentration
effect of residual limiting substrate concentration on
rate of a hypothetical bacterium.
Microbial Growth Kinetics
growth-limiting concentration which will not support
/.L max ' If the organism has a very high affinity for the
limiting substrate (a low K s value) the growth rate will
not be affected until the substrate concentration has
declined to a very low level. Thus, the deceleration
phase for such a culture would be short. However, if
the organism has a low affinity for the substrate (a high
K s value) the growth rate will be deleteriously affected
at a relatively high substrate concentration. Thus, the
deceleration phase for such a culture would be rela-
tively long. Typical values of K, for a range of organ-
isms and substrates are shown in Table 2.2, from which
it may be seen that such values are usually very small
and the affinity for substrate is high. It will be appreci-
ated that the biomass concentration at the end of the
exponential phase is at its highest and, thus, the decline
in substrate concentration will be very rapid so that the
time period during which the substrate concentration is
close to K s is very short.
The stationary phase in batch culture is that point
where the growth rate has declined to zero. However,
as Bull (1974) pointed out, the stationary phase is a
misnomer in terms of the physiology of the organism,
as the population is still metabolically active during this
phase and may produce products called secondary
metabolites, which are not produced during the expo-
nential phase. Bull suggested that this phase be termed
the maximum population phase. The metabolic activity
of the stationary phase has been recognized in the
physiological descriptions of microbial growth pre-
sented by Borrow et al. (1961) and Bu'Lock et al.
(1965). Borrow et al. investigated the biosynthesis of
gibberellic acid by Gibberella jUjikuroi and divided the
growth of the organism into several phases:
(i) The balanced phase; equivalent to the early to
(it) The storage phase; equivalent to the late expo-
nential phase where the increase in mass is due
to the accumulation of lipid and carbohydrate.
TABLE 2.2. Some representative values of K s for a range of
micro-organisms and substrates
Organism Substrate K s (mg dm -3) Referenees
Escherichia coli Glucose 6.8 x 10 - 2 Shehata and
Marr (1971)
Saccharomyces Glucose 25.0 Pirt and
cerevisiae Kurowski (1970)
Pseudomonas sp. Methanol 0.7 Harrison
(1973)
15
 Principles of Fermentation Technology, 2nd Edn.
(iii) The maintenance phase; equivalent to the sta-
tionary phase.
Gibberellic acid (a secondary metabolite) was syn-
thesized only towards the end of the storage phase and
during the maintenance phase. As discussed in Chapter
1, Bu'Lock et al. (1965) coined the terms trophophase,
to refer to the exponential phase, and idiophase to
refer to the stationary phase where secondary
metabolites are produced. The idiophase was depicted
as the period subsequent to the exponential phase in
which secondary metabolites were synthesized. How-
ever, it is now obvious that the culture conditions may
be manipulated to induce secondary metabolism during
logarithmic growth, for example by the use of carbon
sources which support a reduced maximum growth rate
(see Chapter 4).
Pirt (1975) has discussed the kinetics of product
formation by microbial cultures in terms of growth-lin-
ked products and non-growth-linked products.
Growth-linked may be considered equivalent to pri-
mary metabolites which are synthesized by growing
cells and non-growth-linked may be considered equiva-
lent to secondary metabolites. The formation of a
growth-linked product may be described by the equa-
tion:
(2.6)
where p is the concentration of product
and qp is the specific rate of product formation (mg
product g -I biomass h - 1)
Also, product formation is related to biomass produc-
tion by the equation:
dpjdx = I;J!x (2.7)
where I;J!x is the yield of product in terms of biomass
(g product g-I biomass)
Multiply equation (2.7) by dxjdt, then:
dxjdt· dpjdx = I;J!x' dxjdt
and dpjdt = I;J!x' dxjdt.
But dxjdt = p.x and therefore:
dpjdt = I;J!x' p.x
and dpjdt = qp'x
and therefore:
(2.8)
From equation (2.8) it may be seen that when product
16
formation is growth associated the specific
product formation increases with specific
Thus, productivity in batch culture will be greatest
p.rnax and improved product output will be actlle1fed
increasing both p. and biomass concentration.
growth linked product formation is related to bioDl
concentration and, thus, increased productivity in ba
culture should be associated with an increase in
mass. However, it should be remembered that n
growth related secondary metabolites are produced b
under certain physiological conditions primarily
der limitation of a particular substrate so thatt
biomass must be in the correct 'physiological sta
before production can be achieved. The elucidation
the environmental conditions which create the cort
'physiological state' is extremely difficult in batchc
ture and this aspect is developed in a later section.
Thus, batch fermentation may be used to prod
biomass, primary metabolites and secondary me
bolites. For biomass production, cultural conditi
supporting the fastest growth rate and maximum
population would be used; for primary metabolite pt
duction conditions to extend the exponential pha
accompanied by product excretion and for secon
metabolite production, conditions giving a short e
nential phase and an extended production phase,
conditions giving a decreased growth rate in the
phase resulting in earlier secondary metabolite
tion.
CONTINUOUS
Exponential growth in batch culture may be
longed by the addition of fresh medium to the
Provided that the medium has been designed such t
growth is substrate limited (i.e. by some component
the medium), and not toxin limited, exponential gro
will proceed until the additional substrate is exhauste
This exercise may be repeated until the vessel is fu
However, if an overflow device were fitted to the fe
menter such that the added medium displaced an equ
volume of culture from the vessel then continuo
production of cells could be achieved. If medium is f
continuously to such a culture at a suitable rate,
steady state is achieved eventually, that is, formation
new biomass by the culture is balanced by the loss
cells from the vessel. The flow of medium into t
vessel is related to the volume of the vessel by the te
dilution rate, D, defined as:
D =FjV
where F is the flow rate (dm 3 h - 1)

(dm 3 ).
eX1Jresse:o in the units h ~ 1.
in cell concentration over a time
eXj:lresseo as:
= growth - output
dx/dt = J..tX - Dx. (2.10)
conditions the cell concentration
dx/dt = 0 and:
J..tx =Dx (2.11)
J..t=D. (2.12)
ste:aO'v-sl:are conditions the specific growth
by the dilution rate, which is an
variable. It will be recalled that under
conditions an organism will grow at its
growth rate and, therefore, it is
continuous culture may be operated only
below the maximum specific growth
certain limits, the dilution rate may
the growth rate of the culture.
of the cells in a continuous culture of
controlled by the availability of the growth
ch(~mical component of the medium and, thus,
described as a chemostat. The mechanism
the controlling effect of the dilution rate is
relationship expressed in equation (2.5),
l(jh:straeted by Monod in 1942:
J..t = J..tmaxs/(Ks + s).
state, J..t = D, and, therefore,
D = J..tmaxs/(Ks + S)
the steady-state concentration of substrate
the chemostat, and
s= KsD/( J..tmax D). (2.13)
predicts that the substrate concentra-
deltefluirled by the dilution rate. In effect, this
of the cells depleting the substrate to
(;()O¢tmtJratiion that supports the growth rate equal to
rate. If substrate is depleted below the
supports the growth rate dictated by the
the following sequence of events takes
growth rate of the cells will be less than
dilution rate and they will be washed out of
vessel at a rate greater than they are being
resulting in a decrease in biomass
concentfCltion.
substrate concentration in the vessel will
Microbial Growth Kinetics
rise because fewer cells are left in the vessel to
consume it.
(iii) The increased substrate concentration in the
vessel will result in the cells growing at a rate
greater than the dilution rate and biomass con-
centration will increase.
(iv) The steady state will be re-established.
Thus, a chemostat is a nutrient-limited self-balancing
culture system which may be maintained in a steady
state over a wide range of sub-maximum specific growth
rates.
The concentration of cells in the chemostat at steady
state is described by the equation:
(2.14)
where i is the steady-state cell concentration in the
chemostat.
By combining equations (2.13) and (2.14), then:
i=Y[SR-{KsD/(J..tmax D)}]. (2.15)
Thus, the biomass concentration at steady state is
determined by the operational variables, SR and D. If
SR is increased, i will increase but s, the residual
substrate concentration in the chemostat, will remain
the same. If D is increased, J..t will increase (J..t = D)
and the residual substrate at the new steady state
would have increased to support the elevated growth
rate; thus, less substrate will be available to be con-
verted into biomass, resulting in a lower steady state
value.
An alternative type of continuous culture to the
chemostat is the turbidostat, where the concentration
of cells in the culture is kept constant by controlling
the flow of medium such that the turbidity of the
culture is kept within certain, narrow limits. This may
be achieved by monitoring the biomass with a photo-
electric cell and feeding the signal to a pump supplying
medium to the culture such that the pump is switched
on if the biomass exceeds the set point and is switched
off if the biomass falls below the set point. Systems
other than turbidity may be used to monitor the bio-
mass concentration, such as CO 2 concentration or pH
in which case it would be more correct to term the
culture a biostat. The chemostat is the more commonly
used system because it has the advantage over the
biostat of not requiring complex control systems to
maintain a steady state. However, the biostat may be
advantageous in continuous enrichment culture in
avoiding the total washout of the culture in its early
stages and this aspect is discussed in Chapter 3.
The kinetic characteristics of an organism (and,
17
 Principles of Fermentation Technology, 2nd Edn.

therefore, its behaviour in a chemostat) are described

by the numerical values of the constants Y, JL max and
K s • The value of Y affects the steady-state biomass
concentration; the value of JL max affects the maximum ;-------------
/
dilution rate that may be employed and the value of K s /
/

affects the residual substrate concentration (and, hence, /

/

the biomass concentration) and also the maximum dilu- /

tion rate that may be used. Figure 2.4 illustrates the
continuous culture behaviour of a hypothetical bac-
terium with a low K s value for the limiting substrate,
compared with the initial limiting substrate concentra-
tion. With increasing dilution rate, the residual subs-
trate concentration increases only slightly until D ap-
proaches JLmax when s increases significantly. The dilu-
tion rate at which x equals zero (that is, the cells have
been washed out of the system) is termed the critical
dilution rate (DedI) and is given by the equation:
D Cril = JLmaxSR/(Ks
/
Dilution rate
FIG. 2.5. The effect of dilution rate on the steady-state
and residual substrate concentrations in a chemostat of a
organism with a high K s value for the limiting substrate,
pared with the initial substrate concentration.
_ _ _ Steady-state biomass concentration,
+ SR)' (2.16) - - - Steady-state residual substrate concentration,

Thus, DedI is affected by the constants, JL max and K\.,

and the variable, SR; the larger S R the closer is DedI to
JL max . However, JL max cannot be achieved in a simple
steady state chemostat because substrate limited condi- c
o
.;::;
tions must always prevail.
Figure 2.5 illustrates the continuous culture be- ~Cl---~,
~ X at
haviour of a hypothetical bacterium with a high K, for §r-~-
SR2
__ SR3
r----------
the limiting substrate compared with the initial limiting ;;; xat SRI LI SR2_
substrate concentration. With increasing dilution rate, ~r-----_~ I

the residual substrate concentration increases signifi- E

o
! SRI

cantly to support the increased growth rate. Thus, ii

~
there is a gradual increase in s and a decrease in x as t;
D approaches DedI' Figure 2.6 illustrates the effect of >-
"'0
increasing the initial limiting substrate concentration '"
i!l
en ---------
Dilution rate
c FIG. 2.6. The effect of the increased initial substrate
.2
.... tion on the steady-state biomass and residual substrate
~c tions in a chemostat.
OJ
u r---------- _ _ _ Steady-state biomass concentrations.
c I
o - - - Steady-state residual substrate concentrations.
u I
1 SRI, SR2 and SR3 represent increasing concentrations of the
'"'"
'"E I
1 ing substrate in the feed medium.
I
.2 I
.0
~ I

'"
1;:; I
/
on i and s. As SR is increased, so i increases,
>- / residual substrate concentration is unaffected.
"'0 //
'"
OJ ;;/ DedI increases slightly with an increase in SR'
U; -----_ ..... ----- ---- The results of chemostat experiments may
Dilution rate from those predicted by the foregoing theory.
FIG. 2.4'. The effects of dilution rate on the steady-state biomass reasons for these deviations may be anomalies assO
and residual substrate 'concentrations in a chemostat culture of a ated with the equipment or the theory not predict'
micro-organism with a low K s value for the limiting substrate,
c:ompared with the initial substrate concentration.
the behaviour of the organism under certain
_ _ _ Steady-state biomass concentration. stances, Practical anomalies include imperfect
- - - Steadv-state residual substrate concentration. and wall growth. Imperfect mixing would
18

heterogeneity in the fer-
ou(aniSUls being subject to nutrient
under severe limitation. This
paJrticulaxly relevant to very low dilu-
the flow of medium is likely to
pn)blem may be overcome by
'ee(H)aCl\. S~ISH;IU:S, as discussed later in this
another commonly encoun-
which the organism adheres
of the reactor resulting, again, in
helterlog(~ne:lty. The immobilized cells are
rerno'val from the vessel but will consume
l'e$llltmg in the suspended biomass concen-
than predicted. Wall growth may be
the inner surfaces of the vessel with
,...h<,erl1ation in carbon and energy limited
the biomass concentration at low
than predicted. This is attributed
of micro-organisms utilizing a
bl'()p()rtjlon of substrate for maintenance at low
Eflfectiv(;IV, the factor decreases at
Bull has reviewed the major
de\{ialtiOllS from basic chemostat theory.
ch(~m()st,lt may be modified in a number of
COlnmlon modifications are the addi-
stages (vessels) and the feedback of bio-
Multistage systems
tIlUlltis,ta~~e
system is illustrated in Fig. 2.7. The
a multistage chemostat is that different
in the separate stages. This may be
Culture effluent
Culture effluent
.--''--11---, to e ff Iuen t
collection or to
subsequent stages
i----+-i----,
multistage chemostat.
Microbial Growth Kinetics
advantageous in the utilization of multiple carbon
sources and in the production of secondary metabolites.
Harte and Webb (1967) demonstrated that when Kleb-
siella aerogenes was grown on a mixture of glucose and
maltose only the glucose was utilized in the first stage
and maltose in the second. Secondary metabolism may
occur in the second stage of a dual system in which the
second stage acts as a holding tank where the growth
rate is much smaller than that in the first stage. The
adoption of multistage systems in research and industry
has been extremely limited, due to the complexity of
the systems. One example of the industrial application
of the technique is in continuous brewing which is
described in a later section.
Feedback systems
A chemostat incorporating biomass feedback has
been modified such that the biomass in the vessel
reaches a concentration above that possible in a simple
chemostat, that is, greater than Y(S R - s). Biomass
concentration may be achieved by:
(i) Internal feedback. Limiting the exit of biomass
from the chemostat such that the biomass in
the effluent stream is less concentrated than in
the vessel.
(ii) External feedback. Subjecting the effluent
stream to a biomass separation process, such as
sedimentation or centrifugation, and returning
a portion of the concentrated biomass to the
growth vessel.
Pirt (1975) gave a full kinetic description of these
feedback systems and this account summarizes his
analysis.
INTERNAL FEEDBACK
A diagrammatic representation of an internal feed-
back system is shown in Fig. 2.8a. Effluent is removed
from the vessel in two streams, one filtered, resulting in
a dilute effluent stream (and, thus, a concentration of
biomass in the reactor) and one unfiltered. The propor-
tion of the outflow leaving via the filter and the effec-
tiveness of the filter then determines the degree of
feedback. The flow rate of incoming medium is desig-
nated F (dm 3 h -1) and the fraction of the outflow
which is not filtered is designated c; thus the outflow
rate of the unfiltered stream is cF and that of the
19
Principles of Femlentation Technology, 2nd Edn.

(a) If the term 'cO - h) + h' is represented by 'A',

hx Jk =AD.
5
When the filtered effluent stream is cell free then
F1jjf(1.ClF
cF o and A = c. However, if the filter removes
biomass then h will approach I, and when h
there is no feedback and A = h. Thus, the
~_ 5 ~ values of A is c to h and when feedback occurs
1-
less than I, which means that Jk is less than D.
(b) Separator The concentration of the growth limiting sut)str;at~1
in the vessel at steady state is then given by:
F
s= KsAD/( Jkmax - AD)
and the biomass concentration at steady state
by:
x s cF
5 x= Y/A(SR - S).
gx

EXTERNAL FEEDBACK
FIG. 2.8. Diagrammatic representations of chemostats with feed- A diagrammatic representation of an external
back (Pirt, 1975). back system is shown in Fig. 2.Sb. The effluent
(a) Internal feedback. the fermenter is fed through a separator, such
F = flow rate of incoming medium (dm 3 h -1)
continuous centrifuge or filter, which produces
C fraction of the outflow which is not filtered
x = biomass concentration in the vessel and in theunfiltered stream effluent streams - a concentrated biomass stream a
hx = biomass concentration in the filtered stream a dilute one. A fraction of the concentrated stream
(b) External feedback. then returned to the vessel. The flow rate from t
F = flow rate from the medium reservoir (dm 3 h - I) medium reservoir is F (dm3 h - I); the flow rate of
F s = flow rate of the effluent upstream of the separator
x = biomass concentration in the vessel and upstream of the sepa·
effluent upstream of the separator is Fs (dm 3 h -I) a
rator the concentration of biomass in the stream (and in t
hx = biomass concentration in the dilute stream from the separa- fermenter) is x; a is the proportion of the flow whic
tor fed back to the fermenter and g is the factor by wh
g= factor by which the separator concentrates the biomass
the separator concentrates the biomass. Biom
a = proportion of the flow which is fed back to the fermenter
s ~ substrate concentration in the vessel and effluent lines balance in the system will be:
SR substrate concentration in the medium reservoir Change = growth - output + feedback
or dx/dt = JkX - F,.x/V + aFsgx/V.
filtered stream is 0 - c)F. The concentration of the
biomass in the fermenter and in the unfiltered stream The culture outflow (before separation),
is x and the concentration of biomass in the filtered chemostat is:
stream is hx. The biomass balance of the system is: Fs=F+aFs
Change in biomass = Growth - Output in unfil- or Fs=F/(I-a),
tered stream - Output in filtered stream,
substituting F /0 - a) for Fs in equation (2.21)
which may be expressed as: remembering that D = FIV:
dx/dt = JkX - cDx - (I - c)Dhx (2.17) dx/dt = JkX - Dx/(I - a) + agDx/(1 - a).
or: If all the cells are returned to the fermenter
dx/dt = JkX - D{c(1 - h) + h}x. biomass will continue to accumulate in the
However, if the feedback is partial then a steady
At steady state dx/dt = 0, thus: may be achieved, dx/dt = 0 and:
JkX = D{c(1 - h) + h}x Jk = BD
and Jk = D{c(1 - h) + h}. where B = 0 - ag)/O - a).
20

biomass concentrations
are then given by the
~ D), (2.24)
- s). (2.25)
}.L, sand i (2.18, 2.19,
fel'mt~ntl~r with either exter-
be appreciated that:
than growth rate.
in the vessel is in-
bic)m:ass concentration results in
residual substrate compared
than }.L, the critical dilution
rate at which washout occurs)
is applied widely in effluent treat-
the advantages of feedback con-
the process efficiency. The outlet
is considerably less and the
may improve stability in effluent
where mixed substrates of varying
used. The system will also result in
of microbial products as illus-
Bull (1989) who reported very high
PrlJdtlctivities in laboratory biomass recycle
processes are particularly
continuous culture because the ele-
susceptible to oxygen limitation.
seem particularly attractive for
where low growth rates and low cell
pfCJdtlctivity. A number of internal feed-
been developed based on im-
either hollow fibres or microcarriers
7) and it is claimed that with the
in centrifuge design, centrifugal
fe(~dl)ack in suspension cultures may be
1992). The potential for continu-
is considered in the next
Comparison of batch and continuous
culture in industrial processes
of a culture system may be de-
Microbial Growth Kinetics
scribed as the output of biomass per unit time of the
fermentation. Thus, the productivity of a batch culture
may be represented as:
Rbatch = (x max - xo)/(t i tiJ (2.26)
where R is the output of the culture (g biomass
dm- 3 h- J ),
x mo' is the maximum cell concentration
achieved at stationa1y phase,
is the initial cell concentration at inocu-
lation,
is the time during which the organism
grows at }.LmO"
and is the time during which the organism is
not growing at }.Lmax and includes the
lag phase, the deceleration phase and the
periods of batching, sterilizing and har-
vesting.
The productivity of a continuous culture may be repre-
sented as:
R c a n=t · Di(l tiT)
III (2.27)
where R cont is the output of the culture (g biomass
dm -3 h -1),
t iii is the time period prior to the establish-
ment of a steady state and includes ves-
sel preparation, sterilization and opera-
tion in batch culture prior to continuous
operation,
and T is the time period during which steady-
state conditions prevail.
The term Di increases with increasing dilution rate
until it reaches a maximum value, after which any
further increase in D results in a decrease in Di, as
illustrated in Fig. 2.9. Thus, maximum productivity of
biomass may be achieved by the use of the dilution rate
giving the highest value of Di.
I D
erit
Dilution rate
FIG. 2.9. The effect of dilution rate on biomass productivity in
steady-state continuous culture.
21
Principles of Fennentation Technology, 2nd Edn.
The output of a batch fermentation described by
equation (2.26) is an average over the period of the
fermentation and, because the rate of biomass produc-
tion is dependent on initial biomass (dx/dt = /Lx),
the vast proportion of the biomass is produced towards
the end of the fermentation. Thus, productivity in
batch culture is at its maximum only towards the end of
the process. For a continuous culture operating at the
optimum dilution rate, under steady-state conditions,
the productivity will be constant and always maximum.
Thus, the productivity of the continuous system must
be greater than the batch. A continuous system may be
operated for a very long time period (several weeks or
months) so that the negative contribution of the unpro-
ductive time, t iii , to productivity would be minimal.
However, a batch culture may be operated for only a
limited time period so that the negative contribution of
the time, tii , would be very significant, especially when
it is remembered that the batch culture would have to
be re-established many times during the time-course of
a continuous run. Thus, the superior productivity of
biomass by a continuous culture, compared with a
batch culture, is due to the maintenance of maximum
output conditions throughout the fermentation and the
insignificance of the non-productive period associated
with a long-running continuous process.
The steady state achievable in a continuous process
also adds to the advantage of improved biomass pro-
ductivity, as discussed by Hospodka (1966). Cell con-
centration, substrate concentration, product concentra-
tion and toxin concentration should remain constant
throughout the fermentation. Thus, once the culture is
established the demands of the fermentation, in terms
of process control, should be constant. In a batch
fermentation, the demands of the culture vary during
the fermentation - at the beginning, the oxygen de-
mand is low but towards the end the demand is high,
due to the high biomass and the increased viscosity of
the broth. Also, the amount of cooling required will
increase during the process, as will the degree of pH
control. In a continuous process oxygen demand,
cooling requirements and pH control should remain
constant. Thus, the use of continuous culture should
allow for the easier introduction of process automation.
A batch process requires periods of intensive labour
during medium preparation, sterilization, batching and
harvesting but relatively little during the fermentation
itself. However, a continuous process results in a more
constant labour demand in that medium is supplied
continuously sterilized (see Chapter 5), the product is
continuously extracted and the relative time spent on
equipment preparation and sterilization is very small.
22
The argument against continuous biomass
is that the duration of a continuous fermentati
very much longer than a batch one so that there
greater probability of a contaminating organism e
ing the continuous process and a greater probabili
equipment failure. However, problems of contam
tion and equipment reliability are related to equip
design, construction and operation and, prOVided
ficiently rigorous standards are applied, these probl
can be overcome. In fact, the fermentation industry
recognized the superiority of continuous culture for
production of biomass and several large-scale proces
have been established. This aspect is considered
more detail in a later section of this Chapter.
METABOLITE PRODUCTIVITY
Theoretically, a fermentation to produce
metabolite should also be more productive in contin
ous culture than in batch because a continuous cult
may be operated at the dilution rate which mainta
product output at its maximum, whereas in batch
ture product formation may be a transient
menon during the fermentation. The kinetics
formation in continuous culture have been re\'ievved b
Pirt (1975) and Trilli (1990). Product formation
chemostat may be described as:
Change in product concentration = production -
output:
or:
where p is the concentration of product
and qp is the specific rate of product formation
product g- 1 biomass h - I).
At steady state, dp/dt = 0, and thus:
p= qp ·x/D
where p is the steady-state product concentration.
If qp is strictlyrelated to /L, then as D increases so
qp; thus, the steady-state product concentration
and product output (Dp) will behave in the same
as x and Dx, as shown in Fig 2.lOa. If qp is inrlpnen>
dent of /L then it will be unaffected by D and
concentration will decline with increasing D but
will remain constant, as shown in Fig. 2. lOb. If
formation occurs only within a certain range of
rates (dilution rates) then a more complex relatil)ns hll1
is produced.
Thus, from this consideration a chemostat
for the production of a product can be designed
optimize either output (g dm- 3 h- 1) or product
centration. However, as Heijnen et al. (1992) exp,laillea,

D which B could not maintain a t-t of X and would be
Dp
D the production strain will be displaced from the fer-
of D on steady-state product concentration
(Dp ) when:
Q:rclwth-relateet the advantage of high pro-
at high dilution rates must be
the disadvantage of low product con-
res:ulting in increased downstream process-
arguments presented for the supe-
<;oJltiJ1UClUS culture for biomass production
product synthesis ease of automa-
th(~,a,dv<lJ1tagf~s of steady state conditions. The
arises is 'Why has the fermentation
adiClpt:ed continuous culture for the manu-
mi(;rolJial products?' It can be appreciated
arg;uJftents cited previously against continuous
(cc)ntamiJ12ltictn and equipment reliability) are
difficulties have been overcome in
COJltiJ1UClUS biomass processes. The an-
qu,estion lies in the highly selective nature
We have already seen that t-t is
in a steady state chemostat and that
to substrate concentration accord-
Microbial Growth Kinetics
The effect of substrate concentration on specific growth
rate for two organisms, A and B, is shown in Fig. 2.11.
A is capable of growing at a higher specific growth rate
at any substrate concentration. The self-balancing
properties of the chemostat mean that the organism
reduces the substrate concentration to the value where
t-t = D. Thus, at dilution rate X, organism A would
reduce the substrate concentration to Z. However, at
this substrate concentration, organism B could grow
only at a t-t of Y. Therefore, if organisms A and B were
introduced into a chemostat operating at dilution rate
X, A would reduce the substrate concentration to Z at
washed out at a rate of (X - Y) and a monoculture of
A would be established eventually. The same situation
would occur if A and B were mutant strains arising
from the same organism. Commercial organisms have
been selected and mutated to produce metabolites at
very high concentrations (see Chapter 3) and, as a
result, tend to grow inefficiently with low t-tmax values
and, possibly, high K, values. Back mutants of produc-
tion strains produce much lower concentrations of
product and, thus, grow more efficiently. If such back
mutants arise in a chemostat industrial process, then
mentation as described in the foregoing scenario.
Calcott (1981) described this phenomenon as "con-
tamination from within" and this type of 'contamina-
tion' cannot be solved by the design of more 'secure'
fermenters. Thus, it is the problem of strain degenera-
tion which has limited the application of large scale
continuous culture to biomass and, to a lesser extent,
potable and industrial alcohol. The production of al-
cohol by continuous culture is feasible because it is a
byproduct of energy generation and, thus, is not a drain
on the resources of the organism. However, it is possi-
Organism A
~ _ _ Organism B
D
x
y
Residual substrate
concentration
FIG. 2.11. Competition between two organisms in a chemostat.
23
Principles of Fennentation Technology, 2nd Edn.

ble that the technique could be exploited for other
products provided strain degeneration is controlled;
this may be possible in certain genetically engineered
strains. It has been reported that the development of a
continuous process for the manufacture of polyhydroxy-
butyrate, a biopolymer. Other processes have been
developed using chemostat culture.
The adoption of continuous culture for animal cell
products is even more complex than for microbial
systems. Griffiths (1992) compared the following process
options for producing an animal cell product:
is difficult to monitor the genetic stability of
are immobilized in a large reactor system. Thus
scale (kg quantities) animal cell products are stili
duced by batch methods. However, where very
value products are required and production ca
satisfied on a small scale, the continuous per
system is a very attractive proposition (Griffiths,
Continuous brewing and biomass pnJdllction
are the major industrial applications of
microbial culture will now be considered in
tail.

(i) Batch culture.

(ii) Semi-continuous culture where a portion of the
culture is harvested at regular intervals and
replaced by an equal volume of medium.
(iii) Fed-batch culture where medium is fed to the
culture resulting in an increase in volume (see
later section)
(iv) Continuous perfusion where an immobilized
cell population is perfused with fresh medium
and is equivalent to an internal feedback cont-
inuous system.
(v) Continuous culture.
The characteristics of all five modes of operation are
shown in Table 2.3, from which it may be seen that the
perfusion continuous system appears extremely attrac-
tive. However, the practicalities of running a large scale
continuous perfusion system present considerable dif-
ficulties. The process has to be reliable and able to
operate aseptically for the long periods necessary to
exploit the advantage of a continuous process. Also,
the licensing of a continuous process may present some
difficulties where a consignment of product must be
traceable to a batch of raw materials. In a long-term
continuous process several different batches of media
would have to be used which presents the problem of
associating product with raw material. Furthermore, it
CONTINUOUS BREWING
The brewing industry in the United Kingdom
had a relatively brief 'courtship' with continuous
ture. Two types of continuous brewing have been
(i) The cascade or multistage system.
(ii) The tower system.
Hough et al. (1976) described the cascade
lized at Watneys' Mortlake brewery in London.
process utilized three vessels, the first two for
tation and the third for separation of the yeast b
mass. The specific gravity of the wort was reduced fr
1040 to 1019 in the first vessel and from 1019 to 1011
the second vessel. The residence time for the syst
was 15 to 20 hours, using worts in the specific grav'
range of 1035 to 1040, and it could be run continuou
for 3 months. However, it is believed that the syste
was abandoned due to problems of excessive biom
production. This process appears to have been use
widely in New Zealand with greater success (Kirsop
1982), but apparently newer installations are of th
batch type.
A typical tower fermenter for brewing is iIIustrat
in Fig. 2.12. The system is partially closed in

TABLE 2.3. Comparison of the performance of different operational modes of an animal

cell fermentation (After Griffiths, 1992)

Operational Cell No. Product yield Length of run

mode (x 10~6 cm~3) (mg day~l) (mg month~l) (days)

Batch 3 100 200 7

Semi~continuous 3 200 600 21
Fed~batch 6 200 500 14
Perfusion 30 + 3000 12000 > 30
Continuous 2 300 1200 > 100
culture

24

!f~~~~~'JIlBeer outlet
I losses of the process.
I
I Attemporator
I jacket
I from about 1 week to 4 to 8 hours as compared with
I I
I Baffle I Attemporator
~------1
jacket
I I
I I
Sight glass
Sample points
schem:atic representation of a tower fermenter for
(Royston, 1966).
leaves the fermenter due to the
nature of the strains employed. Thus,
a type of internal feed-back. Wort is
the base of the tower and passes
plug of yeast. As the wort rises
it is progressively fermented and
ferme;ntt~r via a yeast-separation zone, which
diameter of the rest of the tower. Hough et
descrilJed the protocol employed in the es-
yeast plug in the tower and the subse-
conditions. Prior to the fermentation,
th()rougll1lv cleaned and steam sterilized,
being more important for the contin-
the batch one. The vessel is filled
wort and inoculated with a labora-
initial stages are designed to encour-
01C)I11:JSS production by the periodical addition
a 9-day period. A porous plug of
\iellelt:lps at the base of the tower. The flow rate
gradually increased over a further 9
time an approximate steady state
Microbial Growth Kinetics
may be achieved. Following the establishment of a high
biomass in the fermenter the system is operated such
that the wort is converted to beer with the formation of
approximately the same amount of yeast as would be
produced in the batch process. The beer produced
during the 3-week start-up period is usually below
specification and would have to be blended with high-
quality beer. Thus, more than 3 months' continuous
operation is necessary to compensate for the initial
The major advantage of the continuous tower process
was that the wort residence time could be reduced
the batch system. However, the development of the
cylindro-conical vessel (initially described by Nathan in
the 1930s, but not introduced until the 1970s, see also
Chapter 7) led to the shortening of the batch fermenta-
tion time to approximately 48 hours. Although this is
still considerably longer than the residence time in a
tower fermenter, it should be remembered that beer
conditioning and packaging takes considerably longer
than the fermentation stage, so that the difference in
the overall processing time between the tower and the
cylindro-conical batch process is not sufficient to justify
the disadvantages of the tower. The major disadvan-
tages of the tower system are the long start-up time,
the technical complexity of the plant, the requirement
for more highly skilled personnel than for a batch
plant, the inflexibility of the system in that a long time
delay would ensue between changing from one beer
fermentation to another and, finally, the difficulty in
matching the flavour of the continuously produced
product with that of the traditional batch product.
Thus, the continuous-tower system has fallen from use,
with virtual universal adoption of the cylindro-conical
batch process.
CONTINUOUS CULTURE AND BIOMASS
PRODUCTION
Microbial biomass which is produced for human or
animal consumption is referred to as single cell protein
(SCP). Although yeast was produced as food on a large
scale in Germany during the First World War (Laskin,
1977) the concept of utilizing microbial biomass as food
was not thoroughly investigated until the 1960s. Since
the 1960s, a large number of industrial companies have
explored the potential of producing SCP from a wide
range of carbon sources. Almost without exception,
these investigations have been based on the use of
continuous culture as the growth technique.
As previously discussed, continuous culture is the
25
Principles of Fermentation Technology, 2nd Edn.
ideal method for the production of microbial biomass.
The superior productivity of the technique, compared
with that of batch culture, may be exploited fully and
the problem of strain degeneration is not as significant
as in the production of microbial metabolites. The
selective pressure in the chemostat would tend to work
to the advantage of the industrialist producing SCP, in
that the most efficient strain of the organism would be
selected, although this is not necessarily the case for
mycelial processes. The development of SCP processes
generated considerable research into large-scale
chemostat design and the behaviour of the production
organism in these very large vessels. Many 'novel' fer-
menters have been designed for SCP processes and
these are considered in more detail in Chapter 7.
A very wide range of carbon sources have been
investigated for the production of SCPo Whey has been
used as a carbon source for biomass production since
the 1940s and such fermentations have been shown to
be economic in that they provide a high-grade feed
product, whilst removing an, otherwise, troublesome
waste product of the cheese industry (Meyrath and
Bayer, 1979). Cellulose has been investigated exten-
sively as a potential carbon source for SCP production
and this work has been reviewed by Callihan et al.
(1979) and Woodward (1987). The major difficulty asso-
ciated with the use of cellulose as a substrate is its
recalcitrant nature.
An enormous amount of research has been con-
ducted into the use of hydrocarbons as sources of
carbon for biomass processes; the hydrocarbons investi-
gated being methane, methanol and n-alkanes. A large
number of commercial firms were involved in this re-
search field but very few created viable, commercial
processes based on SCP production from hydrocarbons
because of the economic difficulties involved (Sharp,
1989). At the start of this research, hydrocarbons were
relatively cheap but, following the 1973 Middle East
War, oil prices escalated and transformed the economic
basis of biomass production from petroleum sources.
ICI were successful in developing a commercial process
for production of bacterial biomass (Pruteen) from
methanol at an annual rate of 54,000 to 70,000 tonnes.
The process utilized a novel air-lift, pressure cycle
fermenter, of 3000 m 3 capacity, and was the first
commercial process to produce SCP from methanol
(King, 1982). The fermentation was run successfully for
periods in excess of 100 days without contamination
(Howells, 1982). Regrettably, the economics of the
process were such that when the price of soya and
fish meal declined Pruteen could not compete as an
26
animal feed. SeIling in bulk ceased in
1989) and the 3000 m 3 vessel was
molished.
The expertise developed by ICI during the
project and RHM's research into the use of a
Fusarium gramineamm, for the production of
food formed the basis of a joint venture Oe1tw(~en
two companies. The ICI pressure cycle PlIIO{-lDJalnf
used to produce the fungal biomass
marketed as Quom) in continuous culture. The
tage of fungal biomass is that it may be processe
give a textured protein which is acceptable for hu
consumption. The low shear properties of the air
vessel conserve the desirable morphology of the fun
The process is operated at a dilution rate of be
0.17 and 0.20 h ~ 1 (/Lmax is 0.28 h ~ 1). The phenome
of mutation and intense selection in the chemostat
proved to be problematical in Myco-protein ferment
tion, because highly branched mutants have arisen
the vessel resulting in the loss of the desirable 1ll
phology. However, the process may stilI be operated
chemostat culture for 1000 hours on the full sc
(Trinci, 1992).
Comparison of batch and continuo
culture as investigative too
Although the use of continuous culture on an indd
trial scale is very limited it is an invaluable investigati
technique. The principle characteristic of batch cult
is change. Even during the log phase cultural COll.
tions are not constant and it is only the const~
maximum specific growth rate which gives the se
blance of stability - biomass concentration, substr
concentration and microbial products all change ex
nentially. During the deceleration phase the onset
nutrient limitation causes the growth rate to decr
from its maximum to zero in a very short time, so it
virtually impossible to study the physiological effects
nutrient limitation in batch culture. As TriIli (19
pointed out, adaptation of an organism to change is
instantaneous, so that the activity of a batch culture
not in equilibrium with the composition of its enviro
ment. Physiological events in a batch culture may ha
been initiated by a change in the environment who
took place some significant time before the change
observed. Thus, it is very difficult to relate 'cause
effect'. The major feature of continuous culture,
other hand, is 'the steady state' - biomass,
and product concentration should remain COllst:ant

growth rate is con-
is nutrient limited.
to exaggerate the signifi-
l-.~An •• ~~ a constant biomass
eCeS~arlly mOlCalll: that the culture is
et a!., 1988). It is possible
rate and other envi-
example temperature, pH
COI1CI~ntrat:iOll. Furthermore, be-
of substrates may be used to
the effects of f.t and
be distinguished.
generate valuable physio-
industrial strain which may
pfiJmi,mtllon of the commercial process.
is the effect of growth rate and
ml~tabollte formation. The inter-
metabolites as compounds pro-
idiopha:se of batch culture may lead one
specific production rates (qp) of
inv'eniely linked to specific growth
of this supposition may be
cp,¢l1lOstat culture and it has been shown to
and thienamycin synthesis
(Lilley et a!., 1981) and
Gil)bel'·ellt1.fitiik~lroi(Bu'Lock et ai., 1974).
relationships have been demon-
:;y:;Ll:ll'L:;. Pirt and Righelato (1967) and
[pstfOdlca (1980) showed that the qp of peni-
correlated with f.t up to a specific
h - 1, after which it is independent
that the apparent negative
related to penicillin degradation.
suggest that the growth rate in a
pel1i<;illiin process should not decline below
cQ1'relatilJns between f.t and qp have
chlortetracycline production by
a/A,reoJracl;ens (Sikyta et a!., 1961), oxytetra-
rimosus (Rhodes, 1984) and
Streptomyces erythraeus (Trilli et at.,
selective nature of the chemostat, which
for industrial production,
tool for the isolation and im-
l111 Icro,-01·ganisms. The use of continuous
is considered in Chapter 3, from
that continuous enrichment cul-
COllSi(:lerabJte advantages over batch enrich-
:pnjcHles and that continuous culture may be
($llGcessfullv to select strains producing higher
microbial enzymes.
Microbial Growth Kinetics
FED·BATCH CULTURE
Yoshida et a!. (1973) introduced the term fed-batch
culture to describe batch cultures which are fed contin-
uously, or sequentially, with medium, without the re-
moval of culture fluid. A fed-batch culture is es-
tablished initially in batch mode and is then fed accord-
ing to one of the following feed strategies:
0) The same medium used to establish the batch
culture is added, resulting in an increase in
volume.
(ii) A solution of the limiting substrate at the same
concentration as that in the initial medium is
added, resulting in an increase in volume.
(iii) A concentrated solution of the limiting subs-
trate is added at a rate less than in (i) and (ii),
resulting in an increase in volume.
(iv) A very concentrated solution of the limiting
substrate is added at a rate less than in 0), (ii)
and (iii), resulting in an insignificant increase
in volume.
Fed-batch systems employing strategies 0) and (ii) are
described as variable volume, whereas a system em-
ploying strategy (iv) is described as fixed volume. The
use of strategy (iii) gives a culture intermediate between
the two extremes of variable and fixed volume.
The kinetics of the two basic types of fed-batch
culture, variable volume and fixed volume, will now be
described.
Variable volume fed-batch culture
The kinetics of variable volume fed-batch culture
have been developed by Dunn and Mor (1975) and Pirt
(1974, 1975, 1979). The following account is based on
that of Pirt (1975). Consider a batch culture in which
growth is limited by the concentration of one substrate;
the biomass at any point in time will be described by
the equation:
x, =x o + Y(SR s) (2.30)
where x, is the biomass concentration after time, t
hours,
and X o is the inoculum concentration.
The final biomass concentration produced when s = 0
may be described as x max and, provided that X o is
small compared with x max :
(2.31)
27
Principles of Fermentation Technology, 2nd Edn.
If, at the time when x = x max ' a medium feed is started
such that the dilution rate is less than Jh max , virtually
all the substrate will be consumed as fast as it enters
the culture, thus:
(2.32)
where F is the flow rate of the medium feed,
and X is the total biomass in the culture, described
by X = xV, where V is the volume of the
culture medium in the vessel at time t.
From equation (2.32) it may be concluded that input
of substrate is equalled by consumption of substrate by
the cells. Thus, (ds j dt) "'" O. Although the total
biomass in the culture (X) increases with time, cell
concentration (x) remains virtually constant, that is
(dxjdt) "'" 0 and therefore Jh "'" D. This situation is
termed a quasi steady state. As time progresses the
dilution rate will decrease as the volume increases and
D will be given the expression:
D = F j (Va + Ft) (2.33)
where Va is the original volume. Thus, according to
Monod kinetics, residual substrate should decrease as
D decreases resulting in an increase in the cell concen-
tration. However, over most of the range of Jh which
will operate in fed-batch culture, SR will be much
larger than K s so that, for all practical purposes, the
change in residual substrate concentration would be
extremely small and may be considered as zero. Thus,
provided that D is less than Jhmax and K s is much
smaller than S R' a quasi steady state may be achieved.
The quasi steady state is illustrated in Fig. 2.13a. The
major difference between the steady state of a chemo-
stat and the quasi steady state of a fed-batch culture is
that Jh is constant in the chemostat but decreases in
the fed-batch.
Pirt (1979) has expressed the change in product
concentration in variable volume fed-batch culture in
the same way as for continuous culture (see equation
2.28):
dpjdt = qpx - Dp.
Thus, product concentration changes according to the
balance between production rate and dilution by the
feed. However, in the genuine steady state of a chemo-
stat, dilution rate and growth rate are constant whereas
in a fed-batch quasi steady state they change over the
time of the fermentation. Product concentration in the
chemostat will reach a steady state, but in a fed-batch
system the profile of the product concentration over
the time of the fermentation will be dependent on the
relationship between qp and Jh (hence D). If qp is
28
(a)
t-------- x
t-------- SN
C:::====.S(GLS)
o
t
(b)
- - - - - - S(GLS)
o
t
FIG. 2.13. Time profiles of fed-batch cultures.
J-L = specific growth rate
x = biomass concentration
S(GLS) = growth limiting substrate
S N = any other substrate than S(GLS)
(a) Variable volume fed-batch culture.
(b) Fixed volume fed-batch culture.
(Pirt, 1979).
strictly growth related then it will change as Jh
with D and, thus, the product concentration
main constant. However, if qp is constant and
dent of Jh, then product concentration will
the start of the cycle when Dp is greater than
will rise with time as D decreases and qpx
greater than Dp. These relationships are shown
2.14a. If qp is related to Jh in a complex
the product concentration will vary according
relationship. Thus, the feed strategy of a
system would be optimized according to the
ship between qp and Jh.
Fixed volume fed-batch
Pirt (1979) described the kinetics of fixed
fed-batch culture as follows. Consider a batch
in which the growth of the process organism
pleted the limiting substrate to a limiting level.
limiting substrate is then added in a COJtlCtmtlrat(~d
such that the broth volume remains almost
then:
dxjdt = GY

(b)
p centration will rise linearly as for biomass. However, if
Time
,nC,Hl!ral1<Jll (p) in fed-batch culture when qp
) or non-growth related, i.e. q p
sul)stituting for dx/dt in equa-
(2.35)
does not exceed JLmax then the
be consumed as soon as it enters
ds/dt == O. However, dx/dt may
zero, as in the case of variable
the biomass concentration,
amount of biomass in the fer-
with time. Biomass concentration
(2.36)
biomass after operating in fed batch
concentration at the onset of
r¢cl-l)at(;h culture.
IlCl'eases then the specific growth rate will
equation (2.35). The behaviour of
recl-l)atc;h culture is illustrated in Fig.
may be seen that JL declines
eCHlatJLOn (2.35)), the limiting substrate
n';l11d.'l1" virtually constant, biomass in-
COllcentl'ation of the non-limiting nutri-
(:iescril)ed the product balance in a fixed
as:
Microbial Growth Kinetics
but substituting for x from equation (2.36) gives:
dp/dt = qp(x" + GYt).
If qp is strictly growth-rate related then product con-
qp is constant then the rate of increase in product
concentration will rise as growth rate declines, i.e. as
time progresses and x increases. These relationships
are shown in Fig. 2.14b. If qp is related to JL in a
complex manner then the product concentration will
vary according to that relationship. As in the case of
variable volume fed-batch the feed profile would be
optimized according to the relationship between qp
and JL.
Cyclic fed-batch culture
The life of a variable volume fed-batch fermentation
may be extended beyond the time it takes to fill the
fermenter by withdrawing a portion of the culture and
using the residual culture as the starting point for a
further fed-batch process. The decrease in volume re-
sults in a significant increase in the dilution rate (as-
suming that the flow rate remains constant) and thus,
eventually, in an increase in the specific growth rate.
The increase in JL is then followed by its gradual
decrease as the quasi steady state is re-established.
Such a cycle may be repeated several times resulting in
a series of fed-batch fermentations. Thus, the organism
would experience a periodic shift-up in growth rate
followed by a gradual shift-down. This periodicity in
growth rate may be achieved in fixed volume fed-batch
systems by diluting the culture when the biomass
reaches a concentration which cannot be maintained
under aerobic conditions. Dilution results in a decline
in x and, thus, according to equation (2.35) an increase
in JL. Subsequently, as feeding continues, the growth
rate will decline gradually as biomass increases and
approaches the maximum sustainable in the vessel once
more, at which point the culture may be diluted again.
Dilution would be achieved by withdrawing culture and
refilling to the original level with sterile water or
medium not containing the feed substrate.
Application of fed-batch culture
The use of fed-batch culture by the fermentation
industry takes advantage of the fact that the concentra-
tion of the limiting substrate may be maintained at a
very low level, thus avoiding the repressive effects of
29
Principles of Fermentation Technology, 2nd Edn.
high substrate concentration. Furthermore, the fed-
batch system also gives some control over the organ-
isms' growth rate, which is also related to the specific
oxygen uptake rate giving some control over the oxygen
demand of the fermentation (see Chapter 9). Both
variable and fixed volume systems result in low limiting
substrate concentrations, but the quasi steady-state of
the variable volume system has the advantage of main-
taining the concentrations of both the biomass and the
non-limiting nutrients constant. Pirt (1979) cites this as
an important feature because the concentrations of
substrates other than those which limit growth can
have a significant effect on biomass composition and
product formation.
The obvious advantage of cyclic fed-batch culture is
that the productive phase of a process may be extended
under controlled conditions. However, a further advan-
tage lies in the controlled periodic shifts in growth rate
which may provide an opportunity to optimize product
synthesis. Dunn and Mor (1975) pointed out that
changes in the rates of chemical processes can give rise
to increases in intermediate concentrations and similar
effects may be possible in microbial systems. This
observation is particularly relevant to secondary
metabolite production which is maximal in batch cul-
ture during the deceleration phase. Bushell (1989) sug-
gested that optimum conditions for secondary
metabolite synthesis may occur during the transition
phase after the withdrawal of a volume of broth from
the vessel and before the re-establishment of the
steady-state following the resumption of the nutrient
feed. During this period the dilution rate will be greater
than the growth rate but, according to Bushell, the rate
of uptake of the growth-limiting substrate should re-
spond immediately to the increased substrate concen-
tration. Thus, an imbalance results between the subs-
trate uptake rate and the specific growth rate - this
imbalance then contributing to a diversion of interme-
diates into secondary metabolism.
The advantages of cyclic fed-batch culture must be
weighed against difficulties inherent in the system. Care
has to be taken in the design of cyclic fed-batch
processes to ensure that toxins do not accumulate to
inhibitory levels and that nutrients other than those
incorporated into the feed medium become limiting
(Queener and Swartz, 1979). Also, a prolonged series
of fed-batch cycles may result in the accumulation of
non-producing or low-producing variants.
The early fed-batch systems that were developed did
not incorporate any form of feedbaek control and re-
lied on the inherent quasi steady state to maintain
process stability. However, the use of concentrated
30
feeds resulting in very high biomass cOllCenti'atiollil
the development of more sophisticated
grammes has necessitated the introduction of
control techniques. In such feedback
mentations a process parameter directly related
organism's physiological state is monitored co
ously by an on-line sensor. The signal generated
sensor is then used in a control loop (see Chapter
control the medium feed rate. Parameters which
been utilized in this way include dissolved Oxygell
centration, pH, effluent gas composition and litil
substrate concentration. Examples of these contro
tems are included in the next section and in Chap
Examples of the use of fed-batch cn
Fed-batch culture was used in the productio
bakers' yeast as early as 1915. It was recognized tta
excess of malt in the medium would lead to too
growth rate resulting in an oxygen demand in exc
that which could be met by the equipment. This
suited in the development of anaerobic conditions
the formation of ethanol at the expense of bio
production (Reed and Nagodawithana, 1991). Thus
organism was grown in an initially weak medi
which additional medium was added at a rate les
the maximum rate at which the organism could us
Thus, this process fulfils the criteria stipulated in l'
(1975) kinetic description for the establishment
quasi steady state, that is, a substrate limited cui
and the use of a feed rate equivalent to a dilution
less than JLmax' It is now recognized that bakers'
is very sensitive to free glucose and respiratory act
may be repressed at a concentration of about 5
dm- 3 (Crabtree, 1929). Thus, a high glucose conce
tion represses respiratory activity as well as giving
to a high growth rate, the oxygen demand of
cannot be met. In modern fed-batch processes for
production the feed of molasses is under strict con
based on the automatic measurement of traces
ethanol in the exhaust gas of the fermenter. Altho
such systems may result in low growth rates, the
mass yield is near the theoretical maximum obtai
(Fiechter, 1982).
It is interesting to note that the production ()
combinant proteins from yeast may be achieved
fed-batch culture techniques very similar to that
oped for the bakers' yeast fermentation. Gu
(1991) reported the production of hepatitis B s
antigen (HBsAg) in a 0.9 dm 3 fed-batch reactor
the feed rate was increased exponentially to

HBsAg production was
strain and good productivity
rtlaint,ammg a high growth rate whilst
below that which
reslpidltOlY activity. Ibba et al. (1993)
tecl-batc:n process for recombinant
cel'eVlSl(le, under the control of a con-
cyclic fed-batch process gave
activity of a continuous fermen-
pr10dlllctivity being due to increased
a genetic explanation of the
offered.
ferme:ntclticm provides an excellent exam-
systems in the production of a
The fermentation may be di-
- the 'rapid-growth' phase,
grows at fLmax' and the 'slow-
'pf(JdUlcti,on' phase. Glucose feeds may be
metabolism of the organism during
the rapid-growth phase an excess
an accumulation of acid and a bio-
UClilCl.HU greater than the aeration capacity
fel·rtl(~nter, whereas glucose starvation may result
nitro~(en in the medium being used as a
resulting in a high pH and inadequate
fonmal:ion (Queener and Swartz, 1979). The
l1utation of hexose may be prevented by the use of
carbohydrate such as lactose in
culture (Matelova, 1976). However, ac-
Queeiller and Swartz, considerable increases
have been achieved by the use of com-
,ontroltled feeding of glucose during the rapid
such that the dissolved oxygen or pH is
within certain limits. Both control parame-
:(;)~:;el1ltially measure the same activity in that both
cOlllce:ntl-ation and pH will fall when glucose is
to an increased respiration rate and the
of organic acids when the respiration
ex,cec~as the aeration capacity of the fermenter.
appear to work well in controlling feed
the rapid-growth phase.
the production phase of the penicillin fer-
j)i(~tit,lticm the feed rates utilized should limit the growth
oxygen consumption such that a high rate of
synthesis is achieved, and sufficient dissolved
available in the medium. The control factor
is normally dissolved oxygen because pH
respOllsi,,'e to the effect of dissolved oxygen on
p¢rlicililin synthesis than on growth. As the fed-batch
proceeds then the total biomass, viscosity and
increase until, eventually, the fermen-
oxygen limited. However, limitation may be
Microbial Growth Kinetics
delayed by reducing the feed rate as the fermentation
progresses and this may be achieved by the use of
computer controlled systems.
Suzuki et al. (1987) developed a pH feedback fed-
batch system for the production of thiostrepton from
Streptomyces laurentii. When glucose was exhausted in
the fermentation the pH rose immediately and this
event was used as the signal for the addition of more
feed which consisted of a concentrated glucose, corn
steep liquor, soy bean meal and mineral mixture. This
process maintained a biomass level of 157 g dm ~3 and
a thiostrepton concentration of 10.5 g dm ~3 with a
productivity nine times that of a conventional batch
culture.
Many enzymes are subject to catabolite repression,
where enzyme synthesis is prevented by the presence of
rapidly utilized carbon sources (Aunstrup et al., 1979).
It is obvious that this phenomenon must be avoided in
enzyme fermentations and fed-batch culture is the ma-
jor technique used to achieve this. Concentrated
medium is fed to the culture such that the carbon
source does not reach the threshold for catabolite
repression. For example, Waki et al. (1982) controlled
the production of cellulase by Trichoderma reesei in
fed-batch culture utilizing CO 2 production as the con-
trol factor and Suzuki et al. (1988) achieved high lipase
production from Pseudomonas fluorescens also using
CO 2 production to control the addition of an oil feed.
Shioya (1990) developed a method for the optimiza-
tion of fed-batch systems based on the relationship
between fL and qp' the product specific production
rate. Once the relationship between the two parame-
ters was established a computer control system was
used to maintain the fed-batch at the optimum specific
growth rate (feed rate). The system was tested for a
number of fermentations including histidine (Brevibac-
terium flavum), acid phosphatase and glutathione (s.
cerevisiae), and lysine (Corynebacterium glutamicum).
Specific growth rate was maintained constant using a
feed-forward control profile which was updated
throughout the fermentation as data were collected.
Very promising results were obtained but difficulties
were experienced in generating sufficiently accurate
data to up-date the control system and maintain the
specific growth rate constant.
REFERENCES
AUNSTRUP, K, ANDRESEN, 0., FALeH, E. A. and NIELSEN, T.
K (1979) Production of microbial enzymes. In Microbial
Technology, Vol. 1 (2nd edition), pp. 282~309 (Editors
31
Principles of Fermentation Technology, 2nd Edn.
Peppler, H. J. and Perlman, D.). Academic Press, New
York.
BORROW, A, JEFFERYS, E. G., KESSEL, R. H. J., LLOYD, E. c.,
LLOYD, P. B. and NIXON, I. S. (1961) Metabolism of
Gibberella fujikuroi in stirred culture. Can. 1. Micro. 7,
227-276.
BULL, A T. (1974) Microbial growth. In Companion to
Biochemistry, Selected Topics for Further Study, pp. 415-442
(Editors Bull, AT., Lagnado, J. R., Thomas, J. O. and
Tipton, K. F.). Longman, London.
Bu'I.OCK, J. D., HAMILTON, D., HULME, M. A., POWELL, A. J.,
SHEPHERD, D., SMALLEY, H. M. and SMITH, G. N. (1965)
Metabolic development and secondary biosynthesis in
Penicillium urticae. Can. 1. Micro. 11, 765-778.
BU'LOCK, J. D., DETROY, R. W., HOSTALEK, Z. and
MUNIM-AI.-SHAKARCHI, A (1974) Regulation of biosyn-
thesis in Gibberella fujikuroi. Trans. Br. Mycol. Soc. 62,
377-389.
BUSHELL, M. E. (1989) The process physiology of secondary
metabolite production. In Microbial Products: New Ap-
proaches. Soc. Gen. Microbiol. Symp. 44, pp. 95-120 (Edi-
tors Baumberg, S., Hunter, I. S. and Rhodes, P. M.).
Cambridge University Press, Cambridge.
CALCOTT, P. H. (1981) The construction and operation of
continuous cultures. In Continuous Culture of Cells, Vol.
1, pp.13-26 (Editor Calcott, P. H.). CRC Press, Boca
Raton.
CALLIHAN, C. D. and CLEMMER, J. E. (1979) Biomass from
cellulosic materials. In Economic Microbiology, Vol. 4,
Microbial Biomass, pp. 208-270 (Editor Rose, A H.) .
Academic Press, London.
CRABTREE, H. G. (1929) Observations on the carbohydrate
metabolism of tumours. Biochem. J. 23, 536-545.
DOSTAI.EK, M., HAGGSTROM, C. and MOLIN, N. (1972) Optimi-
sation of biomass production from methanol. Fermenta-
tion Technology Today, Proc. IV Int. Fermentation Symp.,
pp. 497-511 (Editor Terui, G.).
DUNN, I. J. and MaR, J.-R. (1975) Variable-volume continuous
cultivation. Biotechnol. Bioeng. 17, 1805-1822.
EAGON, R. G. (1961) Pseudomonas natriegens, a marine bac-
terium with a generation time of less than ten minutes. 1.
Bacteriol. 83, 736-737.
FIECHTER, A (1982) Regulatory aspects in yeast metabolism.
In Advances in Biotechnology, 1. Scientific and Engineering
Principles, pp. 261-267 (Editors Moo-Young, M., Robin-
son, C. W. and Vezina, C.). Pergamon Press, Toronto.
Gu, M.B., PARK, M. H. and KIM, D.-I. (1991) Growth rate
control in fed-batch cultures of recombinant Saccha-
romyces cerevisiae producing hepatitis B surface antigen
(HBsAg). Appl. Microbiol. Biotechnol. 35, 46-50.
GRIFFITHS, J. B. (1986) Scaling-up of animal cell cultures. In
Animal Cell Culture, a Practical Approach, pp. 33-70 (Edi-
tor Freshney, R. 1.). IRL Press, Oxford.
GRIFFITHS, J. B. (1992) Animal cell processes - batch or
continuous? 1. Biotechnol. 22, 21-30.
HARRISON, D. E. F. (1973) Studies on the affinity of methanol
32
and methane utilising bacteria for their carbon subs-
trates. 1. Appl. Bacteriol. 36,309-314.
HARTE, M. J. and WEBB, F. C. (1967) Utilisation of mixed
sugars in continuous fermentations. Biotech. Bioeng. 9,
205-221.
HEIJNEN, J. J., TERWISSCHA VAN SCHELTINGA, A H. and
STRAATHOF, A J. (1992) Fundamental bottlenecks in the
application of continuous bioprocesses. 1. Biotechnol. 22,
3-20.
HOSPODKA, J. (1966) Industrial applications of continuous
fermentation. In Theoretical and Methodological Basis of
Continuous Culture of Microorganisms, pp. 493-645 (Edi-
tors Malek, I and Fencl, Z.). Academic Press, New York.
HOUGH, J. S., KEEVIL, C. W., MARIC, V., PHILLISKlRK, G. and
YOUNG, T. w. (1976) Continuous culture in brewing. In
Continuous Culture, 6, Applications and New Fields, pp.
226-238. (Editors Dean, A C. R., Ellwood, D. c., Evans,
C. G. T. and Melling, 1.). Ellis Horwood, Chichester.
HOWELL~, E. R. (1982) Single-cell protein and related tech-
nology. Chem. Indust., 7,508-511.
IBBA, M., BONARIUS, D., KUHLA, J., SMITH, A. and KUENZI, M.
(1993) Mode of cultivation is critical for the optimal
expression of recombinant hirudin by Saccharomyces cere-
visiae. Biotech. Letts. 15 (7), 667-672.
KIRSOP, B. H. (1982) Developments in beer fermentation.
Topics in Enzyme and Fermentation Biotechnology, 6,
79-131 (Editor Wiseman, A). Ellis Horwood, Chichester.
KING, P. P. (1982) Biotechnology: an industrial view. J. Chem.
Tech. Biotechnol. 32, 2-8.
LASKIN, A I. (1977) Single cell protein. Ann. Reports on
Fermentation Processes, 1, 151-180 (Editor Perlman, D.).
Academic Press, New York.
LAVERY, M. (1990) Animal cell fermentation. In Fermentation:
A Practical Approach, pp. 205-220. (Editors McNeil, B.
and Harvey, L. M.). IRL Press, Oxford.
LILLEY, G., CLARK, A E. and LAWRENCE, G. C. (1981) Control
of the production of cephamycin C and thienamycin by
Streptomyces cattleya NRRL 8057. 1. Chem. Tech. Biotech-
nolo 31, 127-134.
MAJOR, N. C. and BULL, A T. (1989) The physiology of lactate
production by Lactobacillus delbreukii in a chemostat with
cell recycle. Biotech. Bioeng. 34, 592-599.
MALEK, I., VOTRUBA, J. and RICICA, J. (1988) The continuality
principle and the role of continuous cultivation of mi-
croorganisms in basic research. In Continuous Culture, pp.
95-104. (Editors Kyslik, P., Dawes, E. A, Krumphanzl, V.
and Novak, M.). Academic Press, London.
MATELOVA, V. (1976) Utilization of carbon sources during
penicillin biosynthesis. Folia Microbiologica, 21, 208-209.
MEYRATH, J. and BAYER, K. (1979) Biomass from whey. In
Economic Microbiology, 4. Microbial Biomass, pp. 208-270.
(Editor Rose, A H.). Academic Press, New York.
MONaD, J. (1942) Recherches sur les Croissances des Cultures
Bacteriennes (2nd edition). Hermann and Cie, Paris.
NATHAN, L. (1930) Improvements in the fermentation and
maturation of beers. 1. Inst. Brewing, 36, 538-550.

I\LFEFlMANN, A. W. (1993) Plant cell cul-
l, Biological Fundamentals,
(Editors Rhem, H.-J. and
of fed batch culture with refer-
fermentation. 1. Appl. Chem.
Prii'1cil1les of Microbe and Cell Cultivation.
Pe'[!-O'all;U culture of microbes. Ann. NY.
of microbial processes: a
Nfi,crmryial Growth Dynamics, pp. 1-16.
K., Bazin, M. J. and Keevil, C. W.).
KtJlwws,a, W. M. (1970) An extension of the
Ch!~m,ostat with feedback of organisms. Its
with a yeast culture. J. Gen.
UH"'W'dU, R. c. (1967) Effect of growth ratc
of penicillin by Penicillium chrysogenum
continuous culture. Appl. Micro., 15,
Formation of the colony in the
{jJh,teiclinium. Aust. 1. BioI. Sci., 12, 53-64.
R. (1979) Penicillins; Biosynthetic
In Economic Microbiology, Vol. 3.
of Metabolism, pp. 35-123 (Editor
Acad,emiC Press, London.
TW. (1991) Yeast Tech-
pp. 413-437. Van Nostrand Rein-
The production of oxytetracycline in
Biotech. Bioeng. 26,382-385.
Tower fermentation of beer. Process
HOSPCIDKA, J. (1980) Quantitative physiology
chrysogenum in penicillin fermentation.
22, 289-298.
Bioprotein Manufacture. A Critical Assess-
HClrwOO,:l, Chichester.
MARR, A. G. (1971) 1. Bacteriol. 107,
Microbial Growth Kinetics
SHIOYA, S. (1990) Optimization and control in fed-batch reac-
tors. Adv. Biochem. Eng.
SIKYTA, B., SLEZAK, J. and HEROLD, H. (1961) Growth of
Streptomyces aureofaciens in continuous culture. Appl.
Micro. 9,233-238.
SUZUKI, T, YAMANE, T and SHIMIZU, S. (1987) Mass produc-
tion of thiostrepton by fed-batch culture of Streptomyces
laurentii with pH-stat modal feeding of multi-substrate.
Appl. Microbiol. Biotechnol. 25,526-531.
Suzma, T., MUSHIGA, Y, YAMANE, T and SHIMIZU, S. (1988).
Mass production of lipase by fed-batch culture of Pseudo-
monas fluorescens. Appl. Microbiol. Biotechnol. 27,
417-422.
TRILL!, A. (1990) Kinetics of secondary metabolite production.
In Microbial Growth Dynamics, pp. 103-126 (Editors,
Poole, R. K., Bazin, M. J. and Keevil, C. W.). IRL Press,
Oxford.
TRILL!, A., CROSSLEY, M. V. and KONTAKOU, M. (1987) Rela-
tion between growth and erythromycin production in
Streptomyces erythraeus. Biotechnol. Lett. 9,765-770
TRINCI, A. P. J. (1969) A kinetic study of growth of Aspergillus
nidulans and other fungi. 1. Gen. Micro. 57, 11-24.
TRINCI, A. P. J. (1974) A study of the kinetics of hyphal
extension and branch initiation of fungal mycelia. 1. Gen.
Micro. 81,225-236.
TRINCI, A. P. J. (1992) Mycoprotein: A twenty year overnight
success story. Mycol. Res. 96(1), 1-13.
WAKl, T, LUGA, K. and ICHIKAWA, K. (1982) Production of
cellulase in fed-batch culture. In Advances in Biotech-
nology, Vol. 1. Scientific and Engineering Principles, pp
359-364 (Editors Moo-Young, M., Robinson, C. W. and
Vezina, C.). Pergamon Press, Toronto.
WOODWARD, J. (1987) Utilisation of cellulose as a fermenta-
tion substrate: problems and potential. In Carbon Subs-
trates in Biotechnology, pp.45-66 (Editors Stowell, J. D.,
Beardsmore, A. J., Keevil, C. W. and Woodward, J. R.).
IRL Press, Oxford.
YOSHIDA, F., YAMANE, T and NAKAMOTO, K. (1973) Fed-batch
hydrocarbon fermentations with colloidal emulsion feed.
Biotech. Bioeng. 15, 257-270.
33